Implications for portfolio construction

Commodities exposure can be a valuable portfolio diversifier and return generator, particularly in times of unexpected inflation. This inaugural issue of Index Capital Insights considers construction of a commodities allocation, given the observation that benchmark methodologies vary considerably and therefore measuring beta is not as straightforward as it is for global equities. Some of the market-defining characteristics employed by index providers include value traded, trading volume, annual production, and others. There is no single answer of how to best measure the commodities market, and therefore market beta depends on the benchmark used. If commodities beta were unambiguously measurable, then constructing a commodities portfolio sleave could be analogous to other asset classes with well specified market defining benchmarks, where the allocation is centered around relatively inexpensive beta and combined with risk mitigators or alpha generators only if they can be reliably selected at a reasonable price. The reality is that constructing a commodities allocation is more complex.

To illustrate, let’s compare two sets of commonly used benchmarks, one in US equities and the other in commodities. In the US stock market, Vanguard offers one ETF that tracks CRSP US Total Market Index (ticker: VTI), and another ETF that tracks the Russell 3000 (ticker: VTHR). As of November 30, 2023, Apple (AAPL) was the largest holding in both ETF’s with respective weights of 6.35% and 6.39%[1]. In fact, the top ten holdings of each ETF were identical in constituency and weighting rank order. Total weight in VTI’s top ten was 27.15% and in VTHR’s top ten was 27.37%. Estimates of AAPL’s market beta are identical to the second decimal place whether we model market returns using the CRSP US Total Market Index or the Russell 3000.

Table 1

Notwithstanding methodological differences, modelling the US stock market is a process with a great degree of homogeneity across index providers. Furthermore, we can be confident that stock markets are represented highly accurately by float-weighted broad-based benchmark indices. For a succinct, nearly incontrovertible exposition on the mechanics of stock market and active management returns, see William Sharpe’s 1991 Financial Analysts Article article, “The Arithmetic of Active Management”.

Commonly used commodities benchmarks stand in marked contrast. The S&P GSCI Index (SPGSCI) is known for being more heavily weighted in energy commodities futures than other benchmarks; a comparison with another popular commodities index, the Bloomberg Commodities Index (BCOM), shows the contrast.

Tables 2a, 2b, 2c

Given the divergence of index composition, it follows that index returns and volatility are similar only by coincidence in specific time periods. This is clear in a comparison of broad commodities financial products across benchmarks and index/active categories. The sample of 5 investments in Table 3 below shows 3 index trackers with 3 different benchmarks, and 2 active funds with their respective benchmarks. Annualized returns vary significantly between the indexed products across all time periods. Absolute returns and Sharpe Ratios of the actively managed investments dominate the index trackers across all time periods.

Table 3

The data above implies that if you are inclined to get commodities exposure via index trackers, you should do your homework on the benchmark as well as the product. When I worked at S&P Dow Jones Indices, we used to say investors should get to know the “index inside” a financial product. This is always good practice when using index trackers. For a commodities allocation, I would advocate not only actively selecting your indexes, but also diversifying across both index trackers and the index/active divide.

Table 4

The correlation matrix in Table 4 shows that commodities investments can enhance diversification for stock/bond portfolios, but future returns distributions and covariance between investments are dynamic and difficult to predict. Therefore, an approach embracing both commodities index trackers and active funds is supported by historical data.

To offer a sense of how this could be used in practice, I did a 5-year back-test with the 7 investments in Table 4. Looking back 5 years has the benefit of including the global market dislocation of early 2020. I compared a traditional 60/40 stocks/bonds allocation with 55/35/10 stocks/bonds/commodities. Within the commodities allocation of 10%, I equally weighted each investment. I simulate quarterly rebalancing back to target weights, and here are the results:

Table 5

For beta exposure to equities, where markets are modelled very well with broad-based cap-weighted indexes and alpha is therefore quite rare, proactively select a benchmark family and index tracking product(s). For commodities exposure, which can be a valuable diversifier and return generator particularly during periods of unexpected inflation, diversify across benchmarks, index trackers, and active strategies. Why? Because, in a nutshell, the notion of commodities beta is more art than science – the polar opposite of straightforwardly modelled markets like public equities.

[1] Source: Vanguard’s website.

[2] The minimum investment for new individual investors in ARCIX is $5 million. But AQR's website says there is no minimum for institutions or accounts with financial advisors.